pediatrician, this information affects decisions regarding whether to resuscitate, what diagnostic tests to order or perform, what treatments to administer, and whether to sustain or withdraw mechanical ventilation and other life support methods. These decisions traditionally have been based on the best estimate of gestational age at the anticipated or actual time of delivery. Extensive literature is available on the likelihood of outcome Cite this article as: Garite TJ, Combs CA, Maurel K, et al. A multicenter prospective study of neonatal outcomes at less than 32 weeks associated with indications for maternal admission and delivery. Am J Obstet Gynecol 2017;217:72.e1-9. 0002-9378/$36.00 ª 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2017.02.043 72.e1 American Journal of Obstetrics & Gynecology JULY 2017 Original Research ajog.org based on gestational age at delivery.1-3 More recently, the Eunice Kennedy Shriver National Institute of Child Health and Human Development has advocated the use of other factors in addition to the estimated gestational age, including birth weight, the baby’s sex, singleton birth, and completion of a course of antenatal corticosteroids.4 Collectively, these data can be used to give best estimates of the likelihood of survival and morbidity in the neonatal period as well as the best information for patient counseling. There have been limited data, however, regarding how the indication for the premature delivery affects the outcome of the neonate at each gestational age. There are 2 reasons this information might be of importance. The indication for admission for impending preterm delivery is known at the time counseling and management decisions are made. Also, there are numerous studies showing that many of the complications associated with the indications for delivery are known to adversely affect neonatal outcome. For example, abruptio placentae, a relatively common reason for preterm delivery, is associated more frequently with hypoxia and acidosis,5-9 which are factors known to affect survival and likelihood of neonatal complications such as intraventricular hemorrhage (IVH), respiratory distress syndrome (RDS), and even necrotizing enterocolitis (NEC).8-13 Preterm premature rupture of membranes (PPROM) is more likely to be associated with infection and inflammation and adverse neonatal outcome than preterm labor or maternal reasons for delivery.14-19 Infection and inflammation are known to be associated with neonatal sepsis, periventricular leukomalacia (PVL), and other complications.19-23 Growthrestricted neonates, especially when associated with evidence of potential acidosis and/or impending fetal death, which often is a reason for delivery, have a greater rate of mortality and the most newborn complications.24 The vast majority of these studies comparing outcomes for the various indications for delivery, however, are retrospective, with limited and oftenunreliable information regarding the indications for delivery. In addition, attempting to compare the effect of one indication to the other between studies are unlikely to be valid because the populations studied usually differ. Furthermore, it often is difficult to accurately define both the indication for admission and the reason for delivery in retrospective studies, especially if databases or neonatal records are used. For example, PPROM may be the indication for admission to the hospital, but the actual indication for proceeding to delivery may be chorioamnionitis, fetal distress, or preterm labor (PTL). Finally, the perinatal complications associated with these various indications leading to premature delivery are most likely in the very preterm gestational ages. For these reasons, we chose to perform a prospective observational study of obstetric admissions leading to preterm delivery of neonates category II.31 Beyond defining RDS, BPD, and sepsis as described previously, the data included were derived from the clinician’s diagnosis in the medical record. Statistical analysis Statistical analyses were performed as follows. Admission and delivery characteristics were tested for statistical significance across the reason for admission groups with the c2 test or Fisher exact test for dichotomous or categorical variables, and the Kruskal-Wallis test for continuous variables. The null hypothesis for the primary outcome is that there is no difference in composite neonatal morbidity between neonates delivered after PTL and after PPROM. The percentage of neonates with the composite neonatal morbidity was calculated for neonates delivered after PTL and PPROM. Differences between the PTL and PPROM groups were tested for statistical significance with the c2 test. Neonatal outcomes also were tested for statistical significance across the reason for admission and reason for delivery groups using the c2 test or Fisher exact test for dichotomous or categorical variables and the Kruskal-Wallis test for continuous variables. A stepwise logistic regression analysis of serious neonatal morbidity was completed. The model included as independent variables all reason for admission variables (using the PTL category as the reference) and prognostic factors for serious neonatal morbidity including all admission characteristic variables, gestational age at delivery, route of delivery, newborn sex, receipt of